Skip to end of metadata
Go to start of metadata

These are the notes for the skype conference call held on Wednesday October 2 2013, on the progress and activities of the Social Network Benchmark task force (10:00-11:00).

Participants:

  • Peter, Renzo (VUA); Orri,Mirko (OGL); Alex (NEO); Andrey (TUM); Norbert, Larri, Arnau (UPC)

Summary:

  • Norbert introduces Arnau Prat as a new team member from UPC. Arnau will lead the technical aspects of DAMA/DEX with the collaboration of Xavi while Norbert will continue the leading the Task Force.
  • Deliverables:
    • First year deliverables have been presented on time.
    • Now it is time to work on the Task Force Report. Peter sets the deadline for the end of October. 
    • Norbert is writing the skeleton of the deliverable based on existing materials in Confluence. Later he will ask for contributions to other members of the Task Force.
  • Choke point coverage and documentation: this was already requested at the last concall on Sep. 18th
    • Mirko will document the relevance of Q2, Q5, Q6, Q8 and Q9, including the choke point coverage
    • Orri will provide a short description of each choke point that has not been documented previously in the TPCTC paper. These are [1.5:1.7], [2.5:2.7], [3.4:3.5], [7.1:7.3]
  • Norbert asks for the validation datasets and parameters for the interactive benchmark, and how to run and publish it in the Task Force Report. Renzo says that the methodology is not still ready and Orri remembers that update queries have not yet been defined in the interactive query set. Peter says that we should focus on having the benchmark running in several platforms plus some analysis on the performance. Peter also says that the goal is not to compare performance between solutions. The final agreement is:
    • ASAP (in a few hours) Orri and Mirko will provide the scale factor, validation parameters and validation results of the 12 interactive queries without variants, plus a brief description on how to run it on BSBM
    • With this validation setup, three vendors (Virtuoso, DEX and Neo4J) will run and validate inside BSBM the benchmark on top of their solutions. The preliminary results will be presented in the next Task Force biweekly concall.
    • During the next two weeks Orri will prepare a document explaining what should be a basic test and which measures should be taken for a preliminary analysis of the interactive queries
    • Up to the end of the month, the three vendors will run a basic set of analysis based on Orri's recommendations. The results will be published also in the Task Force Report.
  • Norbert will send this week the table that Xavi has prepared with the participants in each source code module of the dbgen. After a review by all the Task Force, he will replace the copyrights.

Next concall is scheduled for Wednesday, October 16nd.

 

  • No labels